US20160302639A1

US20160302639A1 - Robotic cleaning device and method for landmark recognition

Info

Publication number: US20160302639A1
Application number: US15/101,235
Authority: US
Inventors: Magnus Lindhé; Anders Haegermarck; Petter Forsberg
Original assignee: Electrolux AB
Current assignee: Electrolux AB
Priority date: 2013-12-19
Filing date: 2013-12-19
Publication date: 2016-10-20
Anticipated expiration: 2033-12-19
Also published as: WO2015090399A1; CN105813526B; US10617271B2; KR20160100951A; CN105813526A; JP2017501474A; KR102137857B1; JP6687286B2; EP3082537A1; EP3082537B1

Abstract

A robotic cleaning device having a main body, a propulsion system configured to move the device across a surface and an obstacle detecting device configured to detect obstacles. The device further includes a processing unit configured to position the robotic cleaning device with respect to the detected obstacles from obstacle detecting device output data, and further to control the propulsion system to move the robotic cleaning device across the surface. The processing unit is configured to identify a landmark and its position from the obstacle detecting device output data, to derive at least one characteristic from the landmark to create and store a generated landmark signature. The processing unit is also configured to compare the generated landmark signature with predetermined landmark signatures and to determine whether the generated landmark signature matches one of the predetermined landmark signatures or not and operate the robotic cleaning device based on the determination.

Description

TECHNICAL FIELD

The invention relates to a robotic vacuum cleaner, which is configured to navigate within a surface to be cleaned by recognizing and attaining landmarks and their characteristics and to a method of landmark recognition.

BACKGROUND

Robotic vacuum cleaners are known in the art. In general robotic vacuum cleaners are equipped with drive arrangement in the form of a motor for moving the cleaner across a surface to be cleaned. The robotic vacuum cleaners are further equipped with intelligence in the form of microprocessor(s) and navigation means for causing an autonomous behaviour such that the robotic vacuum cleaners can freely move around and clean a space in the form of e.g. a room.
In many fields of technology, it is desirable to use robots with an autonomous behaviour such that they can freely move around a space without colliding with obstacles.
As an a example, robotic vacuum cleaners exist in the art with the capability of more or less autonomously vacuum cleaning a room in which furniture such as tables and chairs and other obstacles such as walls and stairs are located. Traditionally, these robotic vacuum cleaners have navigated a room by means of using e.g. ultrasound or light waves or laser beams. Further, the robotic vacuum cleaners typically must be complemented with additional sensors, such as stair sensors, wall-tracking sensors and various transponders to perform accurately. Such sensors are expensive and affect the reliability of the robot.
A large number of prior art robotic vacuum cleaner use a technology referred to as Simultaneous Localization and Mapping (SLAM). SLAM is concerned with the problem of building a map of an unknown environment by a robot while at the same time localizing the robot in the environment using the map. This is often done using a horizontal scanning laser for range measurement, combined with odometry, which provides an approximate position of the robot as measured by the movement of the wheels of the robot.
US 2009/0306822 A1 discloses a robot that scans its environment along a horizontal plane, thereby creating a two dimensional (2D) signature of its environments and landmarks located therein.
US 2002/0091466 discloses a mobile robot with a first camera directed toward the ceiling of a room for recognizing a base mark on the ceiling and a line laser for emitting a linear light beam toward an obstacle, a second camera for recognizing a reflective linear light beam from the obstacle. The line laser emits a beam in the form of straight line extending horizontally in front of the mobile robot. This also creates also a 2D signature of the environment of the robot.
A 2D map or signature of the environment may lead to problems since the robot or robotic vacuum cleaner may not be able to deduce its position from the 2D signatures and thus it may get lost or disoriented.
The use of a base mark on the ceiling and markers on the ceiling in general poses certain disadvantages. First, the robot will need to have two cameras with at least one camera “looking” up towards the ceiling and another camera looking in the direction of movement and thus in the direction of the laser beams from the horizontal line laser, this is expensive and complicates the build up of the robot. Further, the user has to position at least one base mark on the ceiling by using a chair or ladder.
In addition known robotic vacuum cleaners require thus quite a substantial initial effort by the user until they operate smoothly and clean a surface or area autonomously.

SUMMARY

An object of the present invention to provide a robotic cleaning device that is easy to set up, accurate in its navigation and positioning and that operates autonomously from the initial cleaning operation.
A further object of the present invention is to provide a robotic cleaning device that is economic and easy to navigate.
Another object of the present invention is to provide a method of operating a robotic cleaning device, which enhances the efficiency of a cleaning operation by a robotic cleaning device.
The above mentioned objects are solved by a robotic cleaning device and by a method of operating the robotic cleaning device, as claimed in the independent claims.
Disclosed herein is a method operating a robotic cleaning device comprising the steps of:

- recording obstacle detecting device output data of a vicinity of a robotic cleaning device with an obstacle detecting device;
- identifying a landmark and its position from the image and deriving information from the landmark to generate an landmark signature, via a processing unit;
- storing the generated landmark signature;
- comparing the generated landmark signature with predetermined landmark signatures;
- if the generated landmark signature does not match one of the predetermined landmark signatures, determining that the generated landmark signature is a new landmark signature; or
- if the generated landmark position signature matches one of the predetermined landmark position signatures, determining that the generated landmark position signature is known; and
- operating the robotic cleaning device on the basis of the determination.

The method allows the robotic cleaning device to position itself in a new or known environment and to learn its surroundings in a quick manner.
The obstacle detecting device may be embodied in the form of a 3D sensor system comprising a camera device which allows to record images. The images may form the obstacle detecting device output data.
The robotic cleaning device may be a robotic vacuum cleaner or a robotic mop.
The landmark may be a fixed object or obstacle in a building such as a counter, a staircase, a door, a kitchen stove, etc or a movable object or obstacle such as a piece of furniture. From such a landmark a landmark signature may be derived. The landmark signature may be based on any combination of position, shape, orientation or other characteristics of a surface, either for flat surfaces or surfaces having another shape.
If a landmark signature is recognized, the position of the landmark can be used to correct the position estimation of the robotic cleaning device and/or the map representation of the robotic cleaning device, eliminating any positional uncertainty that has been accumulated since the landmark was last observed and recognized.
The 3D sensor system may be a sensor system, which is able to determine the three-dimensional position of an observed point in relation to the robotic cleaning device.
The 3D sensor system may for example be embodied as a radar, an infrared sensor, a camera and vertical lasers combination, a 3D camera device, etc.
The predetermined and generated landmark signatures may comprise a three dimensional (3D) characteristic of the respective landmark.
In an embodiment the method may comprise the step of operating the robotic cleaning device by navigating it on the surface of interest.
The method may further comprise the step of positioning and storing the generated landmark signature in relation to the current position of the robotic cleaning device.
This enables the building of a map, a layout or floor plan of the area to be cleaned. The information and positional data is gathered piece by piece and put together. The robotic cleaning device may thus perform simultaneous localization and mapping.
The generated landmark signature may be predetermined in a database, which is connected to the processing unit.
Information may be derived from the landmark and its images, respectively, relating to 3D surface shape of the landmark.
The derived information may be incorporated in the landmark signature.
Further, information may be derived from the landmark and its images, respectively, said information relating to the position of a mid point of a surface of the landmark.
Such derived information may also be incorporated in the landmark signature and form at least a part of the landmark signature.
In another embodiment information may be derived from the landmark and its images, respectively, relating to a direction and/or angle of a surface normal of a surface of the landmark.
Such derived information may preferably be also incorporated in the landmark signature.
The derived information may further include information related to the size or extension of a landmark.
The information may be derived by the processing unit from the images taken by the obstacle detecting device.
Disclosed herein is further a robotic cleaning device comprising a main body, a propulsion system configured to move the robotic cleaning device across a surface of interest. The robotic cleaning device may further comprise an obstacle detecting device configured to detect obstacles and processing unit configured to position the robotic cleaning device with respect to the detected obstacles from obstacle detecting device output data and further to control the propulsion system to move the robotic cleaning device across a surface of interest. The processing unit is configured to identify a landmark and its position from the obstacle detecting device output data, to derive at least one characteristic from the landmark to create and store a generated landmark signature, the processing unit further being configured to compare the generated landmark signature with predetermined landmark signatures and to determine, whether the generated landmark signature matches one of the predetermined landmark signatures or not and the processing unit further being configured to operate the robotic cleaning device on the basis of the determination.
Such a robotic cleaning device can easily navigate in a surrounding that is new or known. The learning process is short and efficient. In addition the robotic cleaning device is capable to distinguish between movable obstacles and fixed obstacles. Basically the robotic cleaning device is looking for- and identifies fixed obstacles, such as walls, doors, staircases, handrails, etc. in order to navigate.
Movable obstacles are more likely to confuse the localization of the robotic cleaning device, since they are not always necessarily in the same place, whereas fixed obstacles generate “known” or positive inputs and help the robotic cleaning device to navigate.
The 3D sensor system may comprise a camera device configured to record images of a vicinity of the robotic cleaning device and a first and second vertical line lasers configured to illuminate said vicinity of the robotic cleaning device. The processing unit may further be configured to derive positional data from the recorded images. The vertical line lasers improve the quality of the images taken.
It is possible that the robotic cleaning device comprises two vertical line lasers arranged laterally of the obstacle detecting device.
It is noted that the invention relates to all possible combinations of features recited in the claims. Further features of, and advantageous with the present invention will become apparent when studying the appended claims and the following description. Those skilled in the art realize that different features of the present invention can be combined to create embodiments other than those described in the following.
The steps of any method disclosed herein do not have to be performed in the exact order disclosed, unless explicitly stated.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention is now described, by way of example, with reference to the accompanying drawings, in which:

FIG. 1 illustrates schematically a top down view on a robotic cleaning device according to the invention;

FIG. 2 illustrates schematically a front view of a robotic cleaning device according to the invention;

FIG. 3 illustrates a robotic cleaning device scanning a landmark;

FIG. 4 illustrates a robotic cleaning device scanning and recording a landmark, in the specific case a piece of furniture positioned against a wall; and

FIG. 5 illustrates a flow chart according to a method of the invention.

DETAILED DESCRIPTION

The invention will now be described more fully hereinafter with reference to the accompanying drawings, in which certain embodiments of the invention are shown. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided by way of example so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. Like numbers refer to like elements throughout the description.
Referring now to FIGS. 1 and 2 which show a top down view and a front view of a robotic cleaning device 2 according to the invention, the robotic cleaning device 2 may comprise a main body 4, a cleaning portion 10 and a propulsion system 8, 8′ comprising at least one driving wheel 6, in the shown example two driving wheels 6, 6′, two drives, each of them connected to one of the driving wheels 6, 6′.
The propulsion system 8, 8′ may alternatively to the driving wheels 6, 6′ be embodied in the form of crawlers connected to some kind of a drive, wheels as illustrated in the figures or as a hovercraft system.
The robotic cleaning device 2 may further comprise an obstacle detecting device embodied as a 3D sensor system 12 having at least one line laser 16, in the illustrated embodiment two line lasers 16, 18. The line lasers 16, 18 may be embodied in the form of vertical line lasers 16, 18 configured to illuminate an angle of observation of the 3D sensor system 12. A processing unit 20 may be connected to the 3D sensor system 12 or integrally arranged therein and configured to control the propulsion system 8, 8′.
The 3D sensor system 12 may further comprise a camera device 14.
The processing unit 20 may be embodied in the form of one or more microprocessors and it is configured to execute a computer program downloaded to a suitable storage medium 22 associated with the processing unit 14, such as a Random Access Memory (RAM), a Flash memory or a hard disk drive. The processing unit 20 is configured to carry out a method according to embodiments of the present invention when the appropriate computer program comprising computer-executable instructions is downloaded to the storage medium 22 and executed by the processing unit 20. The storage medium 22 may also be a computer program product comprising the computer program. Alternatively, the computer program may be transferred to the storage medium 22 by means of a suitable computer program product, such as a digital versatile disc (DVD), compact disc (CD) or a memory stick. As a further alternative, the computer program may be downloaded to the storage medium 22 over a network. The processing unit 20 may alternatively be embodied in the form of a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a complex programmable logic device (CPLD), etc.
The processing unit 20 may be configured to control a drive controller 24, which is preferably connected to the processing unit 20, as illustrated in FIG. 2. The drive controller 24 may further be configured to control the drives 8, 8′ in order move the robotic cleaning device across a surface 26, 26′ of interest. The surface 26, 26′ of interest may be a surface to clean.
With respect to FIG. 2, for illustrational purposes, the 3D sensor system 12 is separated from the main body 4 of the robotic cleaning device 2. However, in a practical implementation, the 3D sensor system 12 is likely to be integrated with the main body 4 of the robotic cleaning device 2 to minimize the height of the robotic cleaning device 2, thereby allowing it to pass under obstacles, such as e.g. a sofa
The driving wheels 6, 6′ may be configured to be moved independently form each other via the drives 8, 8′, for example via the drive controller 24. Each of the driving wheels 6, 6′ may comprise a drive 8, 8′. Each drive 8, 8′ may further include a suspension and a gear box for the according driving wheel 6, 6′.
The 3D sensor system 12 may be embodied in the form of infrared (IR) sensors and/or sonar sensors, a microwave radar, a camera system registering its surroundings, a 3D camera system, a laser scanner, etc. for detecting obstacles and communicating information about any detected obstacle to the processing unit 20. The processing unit 20 communicates with the drives to control movement of the wheels 8, 8′ in accordance with information provided by the 3D sensor system 12 such that the robotic cleaning device 10 can move as desired across the surface of interest.
The 3D sensor system 12 is configured to take images from a vicinity of the robotic cleaning device 2 while it is moving on a surface 26, 26′ of interest. The 3D sensor system 12 and the vertical line lasers 16, 18 are looking in a forward direction (c.f. FIG. 1) when the robotic cleaning device 2 is moving. The processing unit 20 is configured to process said images, to identify landmarks 28, 28′ within the images and generate or create information, characteristics or signatures therefrom.
The processing unit 20 is configured to identify fixed and unmovable landmarks 28, 28′ in the image and to distinguish these landmarks 28, 28′ from objects that are not fixed such as furniture or the like.
Turning now to FIG. 3, which illustrates exemplary the robotic cleaning device 2 recording a landmark 28 in the form of a wall portion. The robotic cleaning device 2 positions itself in front of the landmark 28 in order to observe it. By attaining images via the 3D sensor system 12 and the camera device 14, respectively, the robotic cleaning device 2 and the processing unit 20, respectively create a landmark signature out of the landmark 28, which in the shown case is a straight, flat wall.
Each of the vertical line lasers 16, 18 illuminates the floor and obstacles in front of the robotic cleaning device 2. The illuminated space or line is observed by the 3D sensor system. The 3D sensor system is positioned with an offset with respect to the vertical plane formed by the vertical line laser 16, 18, which allows it to use the parallax to determine the 3D position of any objector obstacle that reflects the laser beams. This enables the robotic cleaning device to determine a 3D representation of the objects or obstacles illuminated by each vertical line laser 16, 18. For the sake of simplicity only one vertical laser plane of one vertical line laser 16 is illustrate in FIGS. 3 and 4. It falls within the scope of the invention to arrange and use more than one vertical laser plane and thus more than one vertical line laser 16, 18 (c.f. FIG. 2) in order to illuminate the obstacles.
The vertical laser plane has a vertical angle β, said angle chosen so that the vertical laser plane illuminates at least an angle of view of the 3D sensor system 12.
In order to illuminate and record the obstacles and a landmark 28 the robotic cleaning device 2 may be configured to rotate so that the vertical line laser 16 is scanning the whole landmark 28, as indicated in FIG. 4. Alternatively to the robotic cleaning device 2 rotating, it is possible that only the vertical liner lasers 16, 18 are rotating or pivoting.
As illustrated in FIG. 4, the landmark 28′ may be a piece of furniture and may comprise one or more of the following features or characteristics: a location of a midpoint M of the landmark, an orientation and start point of a surface normal N of the landmark, information regarding the horizontal extension D of the landmark, information regarding the size A of a surface of the landmark and/or information regarding the shape of the surface of the landmark. These features can be extracted by translating or rotating, the robotic cleaning device so that vertical line lasers 16, 18 can scan the vicinity and so that the camera device 14 can create a 3D representation of the full volume in front of it. In this 3D representation, landmarks can be identified. One possible class of such landmarks are flat surfaces, such as A.
In FIG. 4 the surface information may for example be “flat” and surface may be defined by coordinates of a congruent plane in the 3D space.
In addition the method and the robotic cleaning device according to the present invention further allow to measure and extract characteristics related to the vertical distance B, which shows elevation of the furniture above ground, the height H of the main corpus of the furniture or the total height C of the furniture.
Extracting the above mentioned features may create a rich landmark signature that is unlikely to match other potential landmark signatures.
From one of the above indicated characteristics for example the horizontal extension D, the processing unit 20 creates or generates a significant, unique landmark signature comprising at least one 3D characteristic of the landmark.
In order to enhance accuracy two or more of the characteristics may be combined in order to create a unique and specific landmark signature.
Optionally it may also be possible to obtain and store information relating to a vertical extension V of the wall. This depends on the room, in case the room has the same vertical extension V in all areas, this may not be of use. However, it is possible that a room has different vertical extensions V or heights and in such a case creating a characteristic relating to this vertical extension V may result in a characteristic that can be used to create a predetermined or generated landmark signature.
The vertical extension V may further be used to calculate and position the midpoint M on the surface of the landmark.
All characteristics mentioned above are stored, for example via coordinates and vectors, within a three dimensional (3D) space. generated and predetermined landmark signatures are thus preferably 3D landmark signatures. The surface normal N for instance, may be characterized by a vector with 3D start and a 3D endpoint (3D coordinates), since the wall illustrated in FIG. 3 may for example be oblique.
The robotic cleaning device 2 and the processing unit 20, respectively, stores the generated landmark signature in relation to its position within the surface 26, 26′ so that during an initial cleaning, positional data of the surface 26, 26′ is created and stored. Once the generated landmark signature is stored for example in a database on the storage medium 22 it becomes a predetermined landmark signature. After at least one landmark signature is generated it is compared with predetermined landmark signatures. During the first initial cleaning operation the robotic cleaning device 2 may store most of the generated landmark signatures in order to build up the database.
The method and the robotic cleaning device according to the invention enhance the feedback an accuracy of positioning of the robotic cleaning device by generating matches, thus by generating feedback to the robotic cleaning device, which feedback relates to the information that a landmark is known so that the robotic cleaning device knows where it is positioned. False matches and thus errors in the localization are reduced. The more features or characteristics a landmark can provide or the more features a landmark allows to be extracted, the better is the positioning and localization of the robotic cleaning device.
FIG. 4 illustrates the robotic cleaning device 2 recording and observing another wall of another exemplary surface 26′. The wall consists of two flat- and one cylinder-surface. Besides being capable of recognizing the characteristics illustrated in FIG. 3, the robotic cleaning device 2 and the processing unit 20, respectively, may be configured to recognize the characteristics or features, such as height H of the main corpus, height B above ground of the main corpus, the horizontal extension D of the furniture, the surface midpoint M and/or the direction of the surface normal N, as shown in FIG. 4. From the above characteristics or features the robotic cleaning device 2 creates the generated and thus later on the predetermined landmark signature via the processing unit 20. After the landmark signature is generated, the robotic cleaning device 2 may continue the cleaning operation or directly attend to another landmark, for example in case it is the initial cleaning operation.
Alternatively, in the case shown in FIG. 4, the robotic cleaning device 2 may change its position to create other generated landmark signatures of the same landmark 28′ so that it is capable of recognizing the landmark 28′ later on when approaching it from another direction. In the presented example in FIG. 4 this could, for example be a depth E of the main corpus of the furniture illustrated in FIG. 4.
The landmark recognition may also occur during the cleaning operation while the robotic cleaning device is moving, thus simultaneous localization and mapping.
FIG. 5 illustrates the method steps for operating the robotic cleaning device. The method steps may comprise the recording SO1 of an image of a vicinity of the robotic cleaning device via detecting camera device 14 thereof; identifying the landmark SO2 in the image and its position within the surface 26, 26′ from the image and the position of the robotic cleaning device and deriving information from the landmark to generate SO3 an generated landmark signature via the processing unit 20. The generated landmark signature is then stored SO4 and compared SO5 with already previously predetermined landmark signatures and it is decided, in the decision maker 30, if the generated landmark signature does not match SO6 one of the predetermined landmark signatures; or if the generated landmark position signature matches SO7 one of the predetermined landmark position signatures. Depending on this outcome it is then determined whether the generated landmark signature is known SO7 or new SO6.
The landmark recognition improves the navigation of the robotic cleaning device 2 on a surface 26, 26′ of interest, which has to be cleaned or which should not be cleaned.
The recognition phase, thus the phase in which the robotic cleaning device 2 is learning the layout, map or floor plan and thus the positional data of the surface 26, 26′ of interest, may be shortened by guiding the robotic cleaning device 2, for example via a remote, to specific landmarks 28, 28′ within the surface 26, 26′ and letting it lean and recognize the landmark 28, 28′. Alternatively, the robotic cleaning device 2 may be configured to move autonomously and learn the layout of the surface 26, 26′.
The generated/predetermined landmark signatures may be labelled in order to assign them to rooms or areas in which they are located. The labelled landmarks may then be tied to instructions such as “clean every time”, “clean once a week”, “clean every day”, “clean only in the evening”, “avoid this area”, etc.
The invention has mainly been described above with reference to a few embodiments. However, as is readily appreciated by a person skilled in the art, other embodiments than the ones disclosed above are equally possible within the scope of the invention, as defined by the appended patent claims.

Claims

1. A method of operating a robotic cleaning device comprising the steps of:

recording an obstacle detecting device output data of a vicinity of the robotic cleaning device with an object detecting device;

identifying a landmark and its position from the image and deriving information from the landmark to generate a landmark signature;

storing the generated landmark signature;

comparing the generated landmark signature with predetermined landmark signatures; and

determining, whether the generated landmark signature matches one of the predetermined landmark signatures or not;

operating the robotic cleaning device on the basis of the determination.

2. The method according to claim 1, wherein the obstacle detecting device output data is in the form of an image recorded by a camera device.

3. The method according to claim 1, wherein the landmark signature is based on a three dimensional (3D) characteristic of the landmark.

4. The method according to claim 1, wherein the operating step includes navigating the robotic cleaning device on the basis of the determination.

5. The method according to claim 1, further comprising the step of positioning and storing the generated landmark signature in a map, based the current position of the robotic cleaning device.

6. The method according to claim 1, further comprising the step of deriving information relating to a 3D surface shape of the landmark.

7. The method according to claim 1, further comprising the step of deriving information relating to the position of a mid point of a surface of the landmark.

8. The method according to claim 7, further comprising the step of deriving information relating to a direction and/or angle of a surface normal of the surface of the landmark.

9. A robotic cleaning device comprising:

a main body;

a propulsion system configured to move the robotic cleaning device across a surface of interest;

an obstacle detecting device configured to detect obstacles;

a processing unit configured to position the robotic cleaning device with respect to the detected obstacles from obstacle detecting device output data and further to control the propulsion system to move the robotic cleaning device across the surface of interest,

characterized in that the processing unit is configured to identify a landmark and its position from the obstacle detecting device output data, to derive at least one characteristic from the landmark to create and store a generated landmark signature, the processing unit further being configured to compare the generated landmark signature with predetermined landmark signatures and to determine, whether the generated landmark signature matches one of the predetermined landmark signatures or not and the processing unit further being configured to operate the robotic cleaning device on the basis of the determination.

10. The robotic cleaning device according to claim 9, wherein the obstacle detecting device comprises 3D sensor system.

11. The robotic cleaning device according to claim 9, wherein the 3D sensor system comprises:

a camera device configured to record images of a vicinity of the robotic cleaning device; and

a first vertical line laser and a second vertical line laser configured to illuminate the vicinity of the robotic cleaning device;

the processing unit further being configured to derive positional data from the recorded images.

12. The robotic cleaning device according to claim 9, wherein the predetermined and generated landmark signatures comprise a three dimensional (3D) characteristic of the respective landmark.

13. The robotic cleaning device according to claim 9, wherein the operation includes the navigation of the robotic cleaning device by the processing unit.

14. The robotic cleaning device according to claim 11, wherein the processing unit is configured to derive information from the recorded image relating to a 3D surface shape of the landmark.

15. The robotic cleaning device according to claim 11, wherein the processing unit is configured to derive information from the recorded image relating to a position of a midpoint of a surface of the landmark.

16. The robotic cleaning device according to claim 15, wherein the processing unit is configured to derive information from the recorded image relating to a surface normal of the surface of the landmark.